James Tompkin

James Tompkin is a visual computing researcher focusing on computer vision, computer graphics, and human-computer interaction. His lab develops techniques for image and video creation, editing, analysis, and interaction, emphasizing image and scene reconstruction from multi-camera systems and complex dynamic scenes. Applications span 2D, multi-view, and VR/AR displays.

His research interests include scene reconstruction, neural rendering, light fields, time-of-flight sensors, and perceptual design for augmented reality. He works on algorithms for depth estimation, view synthesis, and interactive systems.

His publications demonstrate expertise in neural fields, Gaussian splatting, and light field processing for VR/AR applications. Key themes include machine learning for scene flow estimation, efficient rendering of large-scale environments, and improving generative model stability.